i ill11 ll111111 ill ill11 ill11 iiiii iiiii 11111 11111 iiiii iiiii ... ill11 ll111111 ill ill11...

I Ill11 ll111111 Ill Ill11 Ill11 IIIII IIIII 11111 11111 IIIII IIIII 1111111111 1111 Ill1 US006639665B2

(12) United States Patent (io) Patent No.: US 6,639,665 B2 Poole (45) Date of Patent: Oct. 28,2003

(54)

(75)

(73)

MULTISPECTRAL IMAGING SYSTEM FOR CONTAMINANT DETECTION

Inventor:

Assignee: Institute for Technology Development,

Gavin H. Poole, Slidell, LA (US)

Jackson, MS (US)

Subject to any disclaimer, the term of this patent is extended or adjusted under 35 U.S.C. 154(b) by 110 days.

Notice:

Appl. No.: 091779,706

Filed: Feb. 9, 2001

Prior Publication Data

US 200210135760 A1 Sep. 26, 2002

Int. Cl? ................................................... GOlJ 3/00 U.S. C1. ................... 3561300; 3561317; 2501223 R;

2091576 Field of Search ..................... 3561300; 2501223 R;

2091539, 540, 541, 542, 576, 577, 580,

“Sorting Cut Roses with Machine Vision” V. Steinmetz et al, 1994 ASAE, vol. 37(4); 1347-1354. “Multispectral Image Textural Analysis for Poultry Car- casses Inspection” Bosoon Park et al, ASAE, Paper No. 946027 (1994). “An Automated Corn Kernel Inspection System Using Machine Vision” by Bincheng Ni et al, ASAEICSAE Meet- ing Presentation, Paper No. 933032, 1993. “Hyperspectral Imaging For Detecting Bruises in Apples” Renfu Lu et al, Paper 993120 ASAE Meeting, 1999. “Multi-spectral Image Analysis using Neural Network Algorithm for Inspection of Poultry Carcasses”, B. Park et al, 1998 J. Agric. Engng Res. 69, 351-363. “Application of hyperspectral imaging spectrometer systems to industrial inspection” Charles T. Willoughby et al, pps.

“Real-Time Multispectral Image Processing For Poultry Inspection” B. Park et al, ASAE Meeting Presentation, Paper No. 983070, 1998. “Characterizing Multispectral Images of Tumorous, Bruised, Skin-Torn, and Wholesome Poultry Carcasses”, B. Park et al, ASAE, vol. 39(5); 1933-1941 (Jun. 1996).

264-270lSPIE VO. 2599.

(List continued on next page.) 581, 582, 587, 701; 4521177, 178, 179,

184; 1981860.1, 860.2, 860.3, 860.4, 860.5 Primary Examiner-Thong Nguyen Assistant E x a m i n e r k n e l C. Lavarias

(56) References Cited

U.S. PATENT DOCUMENTS

4,239,969 A * 1211980 Haas et al. ................... 378157 4,839,522 A * 611989 Bourgeois et al. ..... 2501455.11 5,428,657 A * 611995 Papanicolopoulos et al. . 378186 5,488,479 A 111996 Williams et al. ............ 3561402 5,621,215 A 411997 Waldroup et al. ........ 2501461.2 5,821,546 A 1011998 Xiao et al. .................. 2501458 5,895,921 A 411999 Waldroup et al. ........ 2501461.2 5,914,247 A 611999 Casey et al. .................. 435134 6,114,699 A 912000 Barton et al. .......... 2501339.09

OTHER PUBLICATIONS

B. Park, Y. R. Chen, M. Ngyuen, “Multi-spectral image analysis using neural network algorithm for inspection of poultry carcasses”, J. Agric. Engng Res., vol. 69, pp. 351-363, 1998.*

(74) Attorney, Agent, or FirmXrowel l & Moring LLP

(57) ABSTRACT

An automated inspection system for detecting digestive contaminants on food items as they are being processed for consumption includes a conveyor for transporting the food items, a light sealed enclosure which surrounds a portion of the conveyor, with a light source and a multispectral or hyperspectral digital imaging camera disposed within the enclosure. Operation of the conveyor, light source and camera are controlled by a central computer unit. Light reflected by the food items within the enclosure is detected in predetermined wavelength bands, and detected intensity values are analyzed to detect the presence of digestive contamination.

23 Claims, 6 Drawing Sheets

,l2d

https://ntrs.nasa.gov/search.jsp?R=20080007045 2018-07-10T04:25:14+00:00Z

US 6,639,665 B2

OTHER PUBLICATIONS

“Multispectral Image Co+ccurrence Matrix Analysis for Poultry Carcasses Inspection”, B. Park et al, ASAE, vol. 39(4); 1485-1491 (Paper No. 94-6027). “Texture Analysis of Aerial Photographs” Ronald Lumia et al, Pattern Recognition vol. 16, No. 1, pp. 39-46, 1983. “Commercial Near-Infrared Reflectance Analyzers”, P.C. Williams et a1 (1987) Commercial Analyzer, pps. 107-141. “Leaf Nitrogen Analysis of Poinsettia (Euphorbia pulcher- rima Will 0.) Using Spectral Properties in Natural and Controlled Lighting” G.E. Meyer et a1,ASAE vol. 8(5); Sep.

“AColor Vision System for Peach Grading” Byron K. Miller et al, ASAE, 1989, pps 1484-1490.

1992 pp 715-722.

“The Nature of Remote Sensing” Chapter 1, Schowengerdt,

“Machine Vision for Color Inspection of Potatoes and Apples” Y. Tao et al, ASAE, vol. 38(5); 1555-1561. “Hyperspectral imaging for safety inspection of food and agricultural products” Renfu Lu et al, SPIE vol. 3544, Nov. 5,1998, pps. 121-133. “Intensified Multispectral Imaging System For Poultry Car- cass Inspection” B. Park et al, ASAE Nov./Dec. 1994, vol. 37, No. 6, pp 1983-1988. “Spectral imaging in biomedicine: A selective overview”, Richard M. Levenson et al, SPIE vol. 3438, Jul. 1998, pps.

R.A. 1997, pp. 1-33.

300-312.

* cited by examiner

U S . Patent Oct. 28,2003 Sheet 1 of 6

101 IMMOBILIZATION

SCALDING t- 102

,+-+1*5 EVISCERATION

+PI06 INSPECTION

US 6,639,665 B2

NO

110 REPROCESS

I CHILL 6-112

l+--jjllJ GRADE

Fig.1

U S . Patent Oct. 28,2003 Sheet 2 of 6 US 6,639,665 B2


I

cu’ f

(D/ f

b-

Ji-J


ILLUMINATE CARCASSES I 5600"

ACQUIRE MULTI SECTRAL I MAGES

TEXTURE ANALYSIS - MEAN, VARIANCE IMAGE SETS

ADD MEAN, 405 VARIANCE VALUES

I 407

JUDGE BAD r"l I 408

JUDGE GOOD 11 I1

I 1

( END

Fig. 4


ILLUMINATE CARCASSES

ACQUIRE MULTl SPECTRAL IMAGES

hi- x 3

CALCULATE SPECTRAL ANGLE MAPPER FUNCTION

1 504 COMBINE RULE FILES

TEXTURE ANALYSIS -c MEAN, VARIANCE

IMAGE SETS 505

ADD MEAN, VARIANCE 1 IMAGESETS

Fig. 5

U S . Patent Oct. 28,2003 Sheet 6 of 6

a b c

Mean Fig. 60

US 6,639,665 B2

a b c

Variance

Fig. 6b

US 6,639,665 B2 1 2

MULTISPECTRAL IMAGING SYSTEM FOR CONTAMINANT DETECTION

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

In view of these deficiencies in the historical visual inspection technique, efforts have been made to develop automated or semiautomated systems for detecting the presence of contaminants in meat, poultry and other food prod-

s ucts during processing. Most such systems utilize a tech- ne u,s, Government has a paid-up license in this inven- nique in which the food item is irradiated with light having

tion and the right in limited circumstances to require the a (for in the uv such that it patent Owner to license others on reasonable terms as causes the emission of fluorescent radiation upon striking provided for by the terms of specific cooperative A ~ ~ ~ ~ - fecal matter or ingesta. Fluorescent light emanating from the merit N ~ , 58-6612-9-023 awarded by the National Aeronau- 10 target food item is then measured and compared with a tics and Space Administration (“NASA”). threshold value. If the light thus gathered exceeds the

threshold, a signal indicative of the presence of fecal contamination or ingesta is generated. Such a system is disclosed, for example in U.S. Pat. Nos. 5,621,215 and

detecting the presence of contaminants in food items during Xiao et processing. More particularly, the invention is directed to the U S . Pat. No. 5,914,247 to CaseY et a1 discloses a fecal detection of ingesta and fecal which may be and ingesta contamination detection system which is based present on poultry carcasses being processed on a conveyor on the Premise that the emission of fluorescent light having system in a poultry processing plant. 20 a wavelength between about 660 and 680 nm is indicative of

of food the presence of ingesta or fecal material. Thus, carcasses borne illnesses in the United States, which resulted in being processed are illuminated with UV or visible light approximately 5,000 deaths. This means that at least one out being between 300 and 6oo nm) and of every four United States residents suffered from the illuminated surface is then examined for the emission of

food products. It is thus apparent that the incidence of food the Of such fluorescence in the borne illnesses is significant even in a highly developed 660-680 nm range is compared with that in the 610-620 nm economy, and that the prevention of such diseases is a matter range as a baseline in order to distinguish fluorescent light of great importance and urgency. The foods most likely to be emissions Of the responsible for the transmission of diseases caused by 30 One object of the Present invention is to Provide an common bacterial pathogens are animal products such as red improved Process and apparatus for detection of ingesta and meats, poultry, eggs, seafood and dairy products. fecal contamination on a food item, which achieves

of meat and poultry products can occur, enhanced accuracy and dependability in positively identify- for example, as a result of exposure to ingesta and fecal ing such contaminants. Ingesta and fecal material are some-

tering, Accordingly, in order to minimize the likelihood of contamination,” which term is used to refer to all content of such contamination, it has been necessary to examine each the digestive tract Or

food itern individually to detect the presence of contami- Another object of the invention is to provide a process and nants, Historically, such inspection has been performed apparatus which can reliably detect such contaminants at a visually by U,S,D,A, inspectors, who examine each indi- 40 speed which is compatible with the rate at which chicken vidual food item as it passes through the processing system. carcasses are Processed on a modern Production line.

Still another object of the invention is to provide a method are placed on a processing line conveyor system for dressing for Processing light reflected from a food item, which and inspection. Typically, such conveyors operate at speeds 45 method Produces a signal that reliably indicates the Presence on the order of 100 per minute, with a six inch of ingesta and fecal contamination, and an apparatus which separation between shackles. Even with two inspectors implements such method. continuously performing such inspection, only about two Yet another object of the invention is to provide a food seconds are allotted for the inspection of each carcass. stuff inspection system having an enclosure which excludes

During these two second inspections, the inspector is 50 ambient light from a Portion of a conveyor system in a food required to check for evidence of eight different diseases as Processing Production line, in which light having a Prese-

chicken was live when placed on the production line, and to which are being inspected. check for evidence of ingesta or fecal contamination. Finally, another object of the invention is to provide an Moreover, during a typical business day operating in two 5s automated food inspection system which can quickly and eight hour shifts, a productive poultry processing plant may accurately identify contaminated food items in a food pro- produce as many as 250,000 processed chickens. cessing line.

It is apparent from this brief description that the historical These and other objects and advantages are achieved by visual inspection of poultry carcasses by human inspectors the method and apparatus according to the present invention is problematic, and that it is poorly suited to the effective 60 in which a portion of a poultry processing conveyor line is detection and elimination of contaminants in modern poultry passed through a light excluding enclosure which excludes processing plants. In particular, it requires the inspectors to ambient light. Inside the enclosure, each carcass is illumi- make a subjective determination repeatedly at intervals of nated by a light source which emits light having a prese- less than two seconds throughout an eight hour shift. Such lected spectral content as the carcass passes before a mul- a system is prone to errors, which can lead to the entry of 65 tispectral digital imaging camera. The camera acquires a contaminated poultry products into the commercial distri- multispectral digital image of each chicken carcass, includ- bution system. ing digital numbers indicative of reflectance values for each

BACKGROUND AND SUMMARY OF THE INVENTION

The present i riven tion relates to an optical system for 15 5,895,921 to Waldroup et a1 and U.S. Pat. No. 5,821,546 to

In 1997, there were 76 million reported

form of illness that could be attributed to the consumption of 25 fluorescent light in the 660 to 680 nm range. In a preferred

material in a food processing plant during or after &ugh- 35 times referred to generically herein as ‘‘digestive

a modern poultry processing chicken

well as for certain quality characteristics, to verify that the lected spectral content can be used to illuminate food items

US 6,639,665 B2 3 4

pixel in an optical sensor array, in each of several preselected DETAILED DESCRIPTION OF THE DRAWINGS frequency bands. FIG. 1 is a schematic diagram which shows the processing

A digital Camera suitable for this Purpose may be, for of chickens in a poultry processing plant. Upon entry into example, a multispectral CCD (charge coupled device) the plant, chickens are initially immobilized, as indicated in camera, in which light collected by the camera lens is split 5 step 101. This includes the process of hanging the chickens into a plurality of spectral components or “optical channels”, on metal stunning them and slaughtering them, each of which is focused on a separate CCD array. TYPicallY, Thereafter, they are scalded as described below, by immer- channel separation is achieved by means of an arrangement sion in hot water (step 102). The feathers are then removed of prism elements and optical filters, although other spectral in step 103 by mechanical ‘‘pickers” which most of resolving devices may be used. Such cameras are known, 10 the feathers. In addition, the birds are often passed in close and are commercially available, for example, from Duncan proximity to an open flame, to burn off any ~yi~op~umes” Technologies, Inc. of Auburn, Calif. (hair-like feathers) which remain.

It is of course possible to achieve similar multispectral are transferred to a separate digital image data by using multiple digital imaging devices, facility, where they are eviscerated (step 105). In this stage, such as CCD cameras, each having its own filter, for 15 the carcasses are inspected by USDA inspectors (step 106) i da t ion of a Preselected wavelength band. In this event, and either passed or rejected in step 107. If they are rejected, however, additional processing capacity is required in order in step 108 they are either condemned as unfit for human to adjust for slightly different viewing angles, to control the consumption (and sent to offal in step 109) or are repro- respective cameras and to assure Proper registration of the cessed in step 110 and returned to the processing line. In step acquired images. 111, the carcasses are passed through a final whole bird

The multispectral digital image signals provided by the 2o washer in which they are rinsed inside and out with high camera system are input to a digital computer via a known pressure water. Finally, in step 112, the birds are chilled to “frame grabber” which assembles the data into respective a temperature below 40” F. by immersion in a water and ice image frame files. These data are then processed by the mixture, and thereafter they can be graded by USDAinspec- computer according to either of two processes, including tors (step 113). differing Processing algorithms and method steps, depend- 25 An important consideration for the purpose of ingesta and ing on the nature of the Production line Processing of the fecal contamination detection according to the invention is chicken carcasses. The end result of such computer analysis the manner in which the chicken carcasses are scalded to is the generation of a “yesino” or “goodbad” determination loosen the feathers, TWO alternatives-known as ‘‘hard for each carcass which passes in front of the CCD camera, 3o scalding” and ‘‘Soft scalding”-are commonly used for this SO that those carcasses on which ingesta or fecal contami- purpose, Hard scalding is performed using water at a tern- nation is detected can be removed automatically from the perature of 55-60’ c . for a period of 30-75 seconds, and is Processing line for additional special washing Procedures to generally more effective in loosening the feathers, but tends eliminate the contaminants. also to remove the cuticle layer of skin that give the chicken

An important feature of the present invention is the 35 a deeper yellow color. Soft scalding on the other hand, is enclosure of a portion of a processing conveyor in a light performed at a temperature of 50-54.5” C. for a period of excluding compartment in order to achieve accurate control 9&120 seconds, and retains the original skin color, but of the spectral content of light which impinges on, and is leaves the feathers more firmly attached. Depending on the reflected by, the chicken carcasses which are inspected customer requirements, either process is suitable. during processing. As discussed in more detail hereinafter, the choice as

Other objects, advantages and novel features of the between these two types of scalding dictates which of two present invention will become apparent from the following alternative processing techniques-including differing pro- detailed description of the invention when considered in cessing algorithms-is used to detect contaminants. conjunction with the accompanying drawings. FIG. 2a is a schematic illustration which shows the

BRIEF DESCRIPTION OF THE DRAWINGS 45 physical layout of the components of the detection system in a light excluding enclosure according to the invention. As

the steps in described above, chickens are placed on a conveyor system in the form of a shackle line, for progressing through the

FIG. 2a is a side view schematic depiction of a portion of dressing and inspection process, A portion of the conveyor

of the multispectral digital contamination detection system indicated by the ioa, T~ facilitate detection of con-

invention; passes through the light excluding enclosure 11. 2b is an end view Of the components and light The conveyor enters the enclosure through an opening

excluding enclosure of FIG. 2a. 5s lla, and travels a circuitous route past a series of baffles l lb , FIG. 3 is a schematic diagram which shows the compo- llc, l l d into a central chamber l le . The purpose of the

nents of the multispectral contamination detection system baffles is to prevent the propagation of ambient light from according to the invention; the exterior of the enclosure into the central chamber l le . A

FIG. 4 is a flow chart which illustrates contaminant comparable set of baffles llf; llg, l l h is also provided at an detection processing according to a first embodiment of the 60 exit opening Il i at the opposite end of the enclosure 11, for invention; the same purpose. In order to assure effective blocking of

FIG. 5 is a flow chart which illustrates contaminant ambient light, the interior surfaces of the enclosure 11, detection processing according to a second embodiment of including baffles lla-lld, llf-llh, are covered with a light the invention; and absorbent material with an extremely low reflectivity, such

which is common to the first and second embodiments of the At approximately the point at which the conveyor line 10 invention. passes into the central chamber l l e , the shackles from which

step 104, the

4o

is a flow chart which processing of poultry in a modern processing plant;

a poultry processing conveyor line with Optical components

arranged in a light according to the

system 10 is shown in FIG. 2 4 and moves in the direction

taminants according to the invention, the conveyor line

FIGS. 6a and 6b illustrates a digital filtering technique 65 as flat black paint.

US 6,639,665 B2 5 6

the carcasses are suspended are swung alternately to the left which provides a predetermined spectral profile. In this case, and to the right as shown in FIG. 2b, so as to form two it has been determined that a light source corrected to 5600K parallel lines lob and ~ O C , which pass in front of four is particularly advantageous. Since ambient light has been multispectral digital imaging cameras 12a-12d. The two excluded from the interior of the enclosure 11, only light lines and lob are separated by an additional baffle Ilk, 5 from the light source is reflected from the target carcass. which extends in a vertical plane parallel to the conveyor. At a midpoint 1Oc along the conveyor line inside the enclosure In step 4022 data representing a multispectral image of 11, between the cameras 124 12b and cameras 12c, 124 in light reflected from the target carcass are generated by the a known manner the shackles are rotated 180” about a digital imaging camera 12 in four wavelengths. In particular, vertical axis, so that each carcass is rotated correspondingly, i o it has been determined that the following four wavelength exposing the opposite side which previously faced the baffle bands are advantageous for this purpose: Ilk. In this way, the carcasses carried along line 10a are imaged on a first side by camera 12a and thereafter on the second side by camera 12c, while those on line 10b are imaged on a first side by camera 12b, and on the other side is by camera 12d. The baffle ilk prevents light from one side of the enclosure from passing to the other. After the imaging is completed the carcasses Pass through the baffles 1lf-11h and out of an exit opening l l i .

previously described, each having a field of view 15a-l5d, as shown in FIG. 2a. Each such camera has associated therewith a light source 13a-13d which illuminates the carcasses with light having a preselected wavelength suit- (AI - (A3 + A41 0%. 1)

illumination areas 14a-14d. A light source that is color corrected to 5600K is advantageous, although it is apparent that other light sources may be used.

Light reflected from each carcass as it passes through an illumination area 14a-14d within the enclosure 11, through one of the respective fields of view 15-15d is collected by the multispectral digital imaging cameras 12a-12d. In order to prevent any of the cameras from detecting light scattered from a carcass illuminated in the field of view in an adjacent camera, additional baffles l lm are provided.

12a-12d generates multispectral image data which is correlated as between and

evaluation of each carcass as discussed hereinafter. FIG. 3 is a schematic block diagram that shows the

components of the detection arrangement according to the invention, in which the camera 12 collects light from a target chicken 16 as it passes through light from the illumination 13 in the camera,s field of view within the light excluding enclosure, as shown in FIG. 2a. Astream 45 in a of resulting digital signals representative of a multispectral image of the target carcass 16 is supplied from the camera

h,=750-830 nm h2=450-500 nm h,=500-535 nm h,=550-585 nm The image data generated in step 402 thus provide four

reflectance values, each represented by a digital number (DN), for each pixel included in the acquired image +ne

The cameras 12a-12d may be CCD cameras of the type 20 such digital number for each of the frequency bands h,-h,. In step 403 these values are combined to generate a value I for each pixel, according to the following algorithm:

I = able for detection purpose, as they move through respective 25 A3(Al + A 2 )

wherein n is an integer constant. Subtraction of the constant n in the numerator as indicated in Equation 1, reduces the

30 amount of background noise by altering the DN values of the specified wavelength hi. The result of the above calculation, which is performed in the computer 18 of FIG. 3, is the creation of an image file having a single DN value for each

The DN values generated in step 403 are then filtered in step 404 by a process referred to as “texture analysis,” which characterizes the image according to areal variations in pixel

image files indicative of the mean and variance for pixels 40 within a moving window of a 3x3 pixel mask, as illustrated

in FIGS, 6a and 6b,

in the texture analysis. For each window of nine pixels (a 3x3 pixel set) a mean value of the DN values is determined

35

Each of the

processed in a computer to generate a binary (yes/no) brightness-that is, DN generate new Output

shows the manner Of calculation Of mean

known manner:

0%. 2) i; XI 12 to a “frame grabber” 17, which organizes the data into individual frames or image files which depict the target

,=I mean= ~

N carcass in each preselected wavelength band. The resulting data files are then input to the computer 18, which process them in a manner described hereinafter in order to generate a “yesino” or “good/bad” signal 18a for each carcass which For example, in FIG. 64 three windows a, b and c are is imaged in this manner. 55 indicated by brackets, with the window a being enclosed by

Asoftware program 19 which is stored in a memory of the a heavy line. The mean of the DN values in window a, also computer 18 contains the necessary algorithm 20 and pro- enclosed by a heavy line, is 4.2. Similarly, the mean value cess steps for analyzing the acquired image data and draw- for window b is 3.2 and for window c is 3.7, as also shown ing a good/bad conclusion as described. It also controls the in the set of mean values. As the window is moved over the camera 12 and the processing line (conveyor 16a) in a 60 whole of the image file, a complete new image file of mean known manner in order to coordinate their operation. values is created. It is of course apparent that different sizes

FIG. 4 is a flow chart which illustrates the steps performed and by the method and apparatus according to a first embodi- The technique for calculation of a new image data set of ment of the invention, which is advantageously used for variance values, shown in FIG. 6b, is similar to that used to detecting ingesta and fecal contamination on carcasses 65 calculate the mean values, as shown in FIG. 6a and which have been hard scalded prior to removal of feathers. described above. The variance values in FIG. 6b, however In a first step 401, the carcass is illuminated by a light source are calculated by the following well known formula:

wherein N is the mmber of data elements xi.

Of windows can be for this Purpose.

US 6,639,665 B2

(4 -XI2 r,r, ,=I Vmance = I=I u = cos-1

N - 1 5 [$ ,=I ?)I:($ ,=I r : )

7

US 6,639,665 B2 9

each of said entrance and said exit has a fixed structural configuration that defines a transport oath for said conveyor into and out of said interior, which fixed structural configuration excludes exterior ambient light from said interior.

2. Apparatus for detecting ingesta or fecal contamination on food items which are processed on a conveyor system, comprising:

a light sealed enclosure which excludes exterior ambient light from an interior thereof, said interior of said enclosure being adapted to surround at least a portion of said conveyor system;

at least one light source disposed in said enclosure, for illuminating each successive food item on said conveyor system with light having a predetermined spectral content;

at least one sensor which senses light reflected from said food items in a plurality of frequency bands, and which outputs signals indicative of reflected light intensity in each of said frequency bands; wherein said enclosure comprises,

walls defining the interior of said light excluding enclosure; and

an entrance and an exit for permitting said conveyor to pass through said interior; and

wherein each of said entrance and said exit has a plurality of interdigitated light blocking baffles, defining a transport path for said conveyor, whereby exterior ambient light is excluded from said interior.

3. Apparatus for detecting contamination on food items

said at least one light source comprises a plurality of light sources arranged sequentially along said conveyor, for illuminating food items on sequential segments of said conveyor; and

said at least one sensor comprises a plurality of sensors, each being paired with a respective one of said light sources for detecting light therefrom that is reflected by a food item; and

said apparatus further comprises a plurality of additional baffles arranged on said walls between adjacent pairs of light sources and sensors for preventing each of said sensors from detecting light scattered from a food item illuminated in a field of view of an adjacent sensor.

4. Apparatus for detecting contamination on food items according to claim 3, wherein said additional baffles are disposed substantially transversely to a direction of travel of food items on said conveyor.

5. Apparatus for detecting contamination on food items according to claim 3, wherein:

said plurality of light sources comprises light sources arranged on opposite sides of said conveyor;

said enclosure further comprises a longitudinal baffle disposed between sensors and cameras on opposite sides of said enclosure, whereby light from light sources on one side of said conveyor is prevented from propagating to and being detected by sensors situated on an opposite side of said conveyor; and

shackles of said conveyor are swung alternately to opposite sides of said longitudinal baffle forming separate lines of food items on both sides of said longitudinal baffle.

6. Conveyor apparatus for processing food items, com-

a conveyor for transporting said food items for process-

according to claim 2, wherein:

prising:

ing;

S

10

1s

20

2s

30

3s

40

4s

so

5s

60

65

10 a light sealed enclosure which excludes exterior ambient

light from an interior thereof, said interior of said enclosure surrounding at least a portion of said conveyor;

at least one light source disposed in said enclosure, for illuminating each successive food item on said conveyor with light having a predetermined spectral content; and

at least one sensor disposed in said enclosure, for sensing light reflected from said food items and outputting signals indicative of reflected light intensity; wherein

said enclosure has an entrance and an exit by which said conveyor passes through said interior; and

each of said entrance and said exit has a fixed structural configuration that defines a transport path for said conveyor into and out of said interior, which fixed structural configuration excludes exterior ambient light from said interior.


said at least one light source comprises a plurality of light sources arranged sequentially along said conveyor, for illuminating food items on sequential segments of said conveyor;




9. Conveyor apparatus for processing food items, comprising:

a conveyor for transporting said food items for processing;

a light sealed enclosure which excludes exterior ambient light from an interior thereof, said interior of said enclosure surrounding at least a portion of said conveyor;

at least one light source disposed in said enclosure, for illuminating each successive food item on said conveyor with light having a predetermined spectral content; and

at least one sensor disposed in said enclosure, for sensing light reflected from said food items and outputting signals indicative of reflected light intensity; wherein


said at least one sensor comprises a plurality of sensors, each being paired with a respective one of said light sources for detecting light therefrom that is reflected by a food item;

said apparatus further comprises a plurality of additional baffles arranged on said walls between adjacent pairs of light sources and sensors for preventing each of said sensors from detecting light scattered from a food item illuminated in a field of view of an adjacent sensor;


US 6,639,665 B2 11




10. In a system for detecting contamination on a food item which is supported and transported on a conveyor under control of a computer, during processing of said food item for consumption, the combination comprising:

a light sealed enclosure which excludes exterior ambient light from an interior thereof, said enclosure having a light excluding entrance and a light excluding exit adapted to accommodate transporting of said food item on said conveyor through said interior, each of said entrance and said exit having a fixed structural configuration that defines a transport oath for said conveyor into and out of said interior which fixed structural configuration excludes exterior ambient light from said interior;

at least one light source disposed in said interior of said enclosure, for illuminating said food item as it passes through said enclosure, with radiation having a predetermined spectral content;

at least one imaging device for acquiring image data indicative of radiation reflected from said food item within said enclosure; and

a computer readable medium encoded with a program for causing said computer to control operation of said light source and said imaging device and movement of said food item through said light sealed enclosure by said conveyor, to acquire image information of said food item as it passes through said light sealed enclosure, and for processing said image information to generate an output which is indicative of contamination on said food item.



said at least one imaging device comprises a plurality of sensors, each being paired with a respective one of said light sources for detecting light therefrom that is reflected by a food item; and

said apparatus further comprises a plurality of additional baffles arranged on said walls between adjacent pairs of light sources and imaging devices for preventing each of said imaging devices from detecting light scattered from a food item illuminated in a field of view of an adjacent imaging device.


13. In a system for detecting contamination on a food item which is supported and transported on a conveyor under control of a computer, during processing of said food item for consumption, the combination comprising:


S

10

1s

20

2s

30

3s

40

4s

so

5s

60

65

12 a light sealed enclosure which excludes exterior ambient

light from an interior thereof, said enclosure having a light excluding entrance and a light excluding exit adapted to accommodate transporting of said food item on said conveyor through said interior, each of said entrance and said exit having a fixed structural configuration that defines a transport oath for said conveyor into and out of said interior, which fixed structural configuration excludes exterior ambient light from said interior;

at least one light source disposed in said interior of said enclosure, for illuminating said food item as it passes through said enclosure, with radiation having a predetermined spectral content;


a computer readable medium encoded with a program for causing said computer to control operation of said light source and said imaging device and movement of said food item through said light sealed enclosure by said conveyor, to acquire image information of said food item as it passes through said light sealed enclosure, and for processing said image information to generate an output which is indicative of contamination on said food item, wherein: said at least one light source comprises a plurality of

light sources arranged sequentially along said conveyor, for illuminating food items on sequential segments of said conveyor; and

said at least one imaging device comprises a plurality of sensors, each being paired with a respective one of said light sources for detecting light therefrom that is reflected by a food item;

said apparatus further comprises a plurality of additional baffles arranged on said walls between adjacent pairs of light sources and imaging devices for preventing each of said imaging devices from detecting light scattered from a food item illuminated in a field of view of an adjacent imaging device;

said additional baffles are disposed substantially transversely to a direction of travel of food items on said conveyor;



shackles of said conveyor are swung alternately to opposite sides of said longitudinal baffle forming separate lines of food items on both sides of said longitudinal baffle.

14. The system according to claim 10, wherein said output is a processed image which shows a visual indication of contamination.

15. The system according to claim 10, wherein said computer automatically detects presence of contamination based on said output.

16. The system according to claim 10, wherein said contamination is digestive contamination.

17. In a system for detecting contamination on a food item which is supported and transported on a conveyor under control of a computer, during processing thereof for consumption, the combination comprising:

US 6,639,665 B2 13

a light sealed enclosure which excludes exterior ambient light from an interior thereof, said enclosure having a light excluding entrance and a light excluding exit adapted to accommodate transporting of said food item through said interior, each of said entrance and said exit having a fixed structural configuration that defines a transport oath for said conveyor into and out of said interior, which fixed structural configuration excludes exterior ambient light from said interior;

a conveyor for supporting and transporting said food item through said light sealed inclosure while it is being processed for consumption;

at least one light source for illuminating said food item in said interior of said enclosure, with radiation having a predetermined spectral content;


a computer readable medium encoded with a program for causing said computer to control operation of said conveyor, said light source and said imaging device to acquire image information of said food item as it passes through said light sealed enclosure, for generating an output which is indicative of contamination on said food item.

18. Apparatus for detecting contamination of food item





20. In a system for detecting contamination on a food item during processing thereof for consumption, the combination comprising:

a light sealed enclosure which excludes exterior ambient light from an interior thereof, said enclosure having a light excluding entrance and a light excluding exit adapted to accommodate transporting of said food item through said interior;


S

10

1s

20

2s

30

3s

40

4s

14 a conveyor for supporting and transporting said food item

through said light sealed enclosure while it is being processed for consumption;

at least one light source for illuminating said food item in said interior of said enclosure, with radiation having a predetermined spectral


a computer readable medium encoded with a program for causing said computer to control operation of said conveyor, said light source and said imaging device to acquire image information of said food item as it passes through said light sealed enclosure, for generating an output which is indicative of contamination on said food item; wherein,



said apparatus further comprises a plurality of additional baffles arranged on said walls between adjacent pairs of light sources and sensors for preventing each of said sensors from detecting light scattered from a food item illuminated in a field of view of an adjacent sensor;




21. The svstem according to claim 20. wherein said outuut - is a processed image which shows a visual indication of contamination.

22. The system according to claim 20, wherein said computer automatically detects presence of contamination

23. The system according to claim 20, wherein said so based on said output.

contamination is digestive contamination.

* * * * *

i ill11 ll111111 ill ill11 ill11 iiiii iiiii 11111 11111 iiiii iiiii ... ill11 ll111111 ill ill11...

Documents